Benchmark RadarAI BENCHMARK PROFILE
MM-Snowball
MM-Snowball is a benchmark for diagnosing hallucination snowballing in multimodal multi-turn dialogue, with fine-grained analysis. It includes data and code via a project page.
- Released
- 2026-05-30
- Readiness
- Inspectable
- Primary field
- Cybersecurity
Why it matters
Addresses the lack of benchmarks for error propagation in long-horizon interactions, and shows existing mitigation methods are ineffective, motivating new approaches.
Motivation
Multimodal large language models (MLLMs) demonstrate remarkable visual understanding, yet their reliability in interactive settings is severely undermined by hallucination snowballing: a phenomenon where initial errors amplify across conversational turns, leading to a collapse in coherence.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.